SOFT COMPUTING TECHNIQUES FOR PREDICTION OF BOILING HEAT TRANSFER COEFFICIENT OF LIQUIDS ON COPPER-COATED TUBES

In this article pool boiling heat transfer coefficient (HTC) of liquids (isopropanol, methanol, and distilled water) on copper-coated heating tubes over a wide range of pressure conditions is computed experimentally. The objective is to find the applicability of soft computing techniques, swarm-intelligence based neural network, and adaptive fuzzy models in the prediction of boiling HTC. The results are compared with those computed experimentally. The performance of models for prediction of HTC is analyzed in terms of root mean square of prediction error. The minimum/maximum value obtained by zero-order fuzzy model with six membership function is 0.0023/3.4383 among all the liquids considered. The model is found to predict HTC with a maximum error of ±0.5% for boiling of liquids over all the coated tubes with pressure varying from atmospheric to subatmospheric levels. The study shows an excellent agreement between the experimental and predicted data.